The Prediction of Spark-Ignition Engine Performance and Emissions Based on the SVR Algorithm

Abstract

Engine development needs to reduce costs and time. As the current main development methods, 1D simulation has the limitations of low accuracy, and 3D simulation is a long, time-consuming task. Therefore, this study aims to verify the applicability of the machine learning (ML) method in the prediction of engine efficiency and emission performance. The support vector regression (SVR) algorithm was chosen for this paper. By the selection of kernel functions and hyperparameters sets, the relationship between the operation parameters of a spark-ignition (SI) engine and its economic and emissions characteristics was established. The trained SVR algorithm can predict fuel consumption rate, unburned hydrocarbon (HC), carbon monoxide (CO), and nitrogen oxide (NOx) emissions. The determination coefficient (R2) of experimental measured data and model predictions was close to 1, and the root-mean-squared error (RMSE) is close to zero. Additionally, the SVR model captured the corresponding trend of the engine with the input, though some existed small errors. In conclusion, these results indicated that the SVR model was suitable for the applications studied in this research.

Estimation of the flow rate of pyrolysis gasoline, ethylene, and propylene in an industrial olefin plant using machine learning approaches

“…SVR has already received a lot of interest because of its ability to represent complicated structures. The basic principle of the SVR model is concisely given in this paper 31 , 32 .…”

Optimizing soybean biofuel blends for sustainable urban medium-duty commercial vehicles in India: an AI-driven approach

“…The predictions were based on analysing scalar field contours within the cylinder at the exhaust valve opening phase. The scalar field examining this work includes equivalence ratio, temperature, velocity and turbulent energy Zhang et al ( 2022c ) Multistate deep reinforcement learning (M-DRL) This study presented a novel energy management strategy based on M-DRL with a hybrid action space combining discrete and continuous elements. The state space is expanded to integrate real-time multivariate traffic and terrain information, enhancing the accuracy of the energy management system Li et al ( 2007 ) Gaussian process regression (GPR) GPR model demonstrated exceptional accuracy in predicting performance, temperature and emission values under both steady-state and transient operating conditions, when compared to other regression modeling methods Wong et al ( 2015 ) Sparse Bayesian extreme learning machine (SBELM) In relation to the duration of execution, the size of the model, and its ability to withstand fluctuations in the quantity of hidden neurons, SBELM demonstrates superiority over other approaches.…”

Integrating Machine Learning for Predicting Internal Combustion Engine Performance and Segment-Based CO2 Emissions Across Urban and Rural Settings

“…Numerous researchers have employed machine learning methodologies for the anticipation of engine-related parameters. According to Zhang et al [24], the utilization of support vector regression algorithm in numerical simulations has proven to be highly effective in forecasting engine performance and emissions. Karunamurthy et al [25] conducted a comprehensive review of various machine learning methodologies and algorithms, including artificial neural networks, random forest, semi supervised fuzzy and support vector machine, as utilized by multiple researchers.…”